- Easy installation, few dependancies. No database daemon, web server, desktop environment or exotic language necessary.
- Will run on most Unix-based systems
- Qt 4 GUI, plus command line, Unity Lens, KIO and krunner interfaces.
- Searches most common document types, emails and their attachments. Transparently handles decompression (gzip, bzip2).
- Powerful query facilities, with boolean searches, phrases, proximity, wildcards, filter on file types and directory tree.
- Multi-language and multi-character set with Unicode based internals.
- Extensive documentation, with a complete user manual and manual pages for each command.
Recoll has been compiled and tested on Linux, MacOS X and Solaris (initial versions Redhat 7, Fedora Core 5, Suse 10, Gentoo, Debian 3.1, Solaris 8). It should compile and run on all subsequent releases of these systems and probably a few others too.
Qt versions from 3.1 to 4.8
Recoll can index many document types (along with their compressed versions). Some types are handled internally (no external application needed). Other types need a separate application to be installed to extract the text. Types that only need very common utilities (awk/sed/groff/Python etc.) are listed in the native section.
File types indexed natively
- maildir and mailbox ( Mozilla, Thunderbird and Evolution mail ok).
- gaim and purple log files.
- Scribus files.
- Man pages (needs groff).
- Dia diagrams.
- Excel and Powerpoint for Recoll versions 1.19.12 and later.
File types indexed with external helpers
Many document types need the iconv command in addition to the applications specifically listed.
The XML ones
The following types need xsltproc from the libxslt package. Quite a few also need unzip:
- Abiword files.
- Fb2 ebooks.
- Kword files.
- Microsoft Office Open XML files.
- OpenOffice files.
- SVG files.
- Gnumeric files.
- Okular annotations files.
The following need miscellaneous helper programs to decode the internal formats.
- pdf with the pdftotext command, which used to be installed as part of xpdf but is nowadays more often found with poppler, depending on your distribution (the package name is quite often poppler-utils).
- msword with antiword. It is also useful to have wvWare installed as it may be be used as a fallback for some files which antiword does not handle.
- Wordperfect with the wpd2html command from libwpd. On some distributions, the command may come with a package named libwpd-tools or such, not the base libwpd package.
- Lyx files (needs Lyx to be installed).
- Powerpoint and Excel with the catdoc utilities up to recoll 1.19.12. Recoll 1.19.12 and later use internal Python filters for Excel and Powerpoint, and catdoc is not needed at all (catdoc did not work on many semi-recent Excel and Powerpoint files).
- CHM (Microsoft help) files with Python, pychm and chmlib.
- GNU info files with Python and the info command.
- EPUB files with Python and this Python epub decoding module.
- Tar archives (needs Python). Tar file indexing is disabled
by default (because tar archives don't typically contain the
kind of documents that people search for), you will need to
enable it explicitely, like with the following in your
[index] application/x-tar = execm rcltar
- Zip archives (needs Python).
- Rar archives (needs Python), the rarfile Python module and the unrar utility.
- iCalendar(.ics) files (needs Python, icalendar).
- Mozilla calendar data See the wiki about this.
- postscript with
ghostscript and pstotext.
Pstotext 1.9 has a serious issue with special characters in
file names, and you should either use the version packaged for
your system which is probably patched, or apply the Debian
patch which is stored here for
convenience. See http://packages.debian.org/squeeze/pstotext
To make things a bit easier, I also store an already patched version. I added an install target to the Makefile... This installs to /usr/local, use make install PREFIX=/usr to change. So all you need is:
tar xvzf pstotext-1.9-patched.tar.gz cd pstotext-1.9-patched make make install
- RTF files with unrtf. Please note that up to version 0.21.3, unrtf mostly does not work with non western-european character sets. If you have a need for indexing, e.g., russian or chinese RTF files, make sure that you have 0.21.3 or newer. Unrtf is easy to build from source.
- TeX with untex. If there is no untex package for your distribution, a source package is stored on this site (as untex has no obvious home). Will also work with detex if this is installed.
- dvi with dvips.
- djvu with DjVuLibre.
- Audio file tags. Recoll releases 1.14 and later use a Python filter based on mutagen for all audio types.
- Image file tags with exiftool. This is a perl program, so you also need perl on the system. This works with about any possible image file and tag format (jpg, png, tiff, gif etc.).
- Midi karaoke files with Python, the midi module, and some help from chardet. There is probably a python-chardet package for your distribution, but you will quite probably need to build the midi package. This is easy but see the notes here.
- Konqueror webarchive format with Python (uses the tarfile standard library module).
- Mimehtml web archive format (support based on the mail filter, which introduces some mild weirdness, but still usable).
- Can use a Firefox extension to index visited Web pages history. See the Wiki for more detail.
- Processes all email attachments, and more generally any realistic level of container imbrication (the "msword attachment to a message inside a mailbox in a zip" thingy...) .
- Multiple selectable databases.
- Powerful query facilities, with boolean searches, phrases, filter on file types and directory tree.
- Xesam-compatible query language.
- Wildcard searches (with a specific and faster function for file names).
- Support for multiple charsets. Internal processing and storage uses Unicode UTF-8.
- Stemming performed at query time (can switch stemming language after indexing).
- Easy installation. No database daemon, web server or exotic language necessary.
- An indexer which runs either as a batch, cron'able program, or as a real-time indexing daemon, depending on preference.
The Recoll GUI has many features that help to specify an efficient search and to manage the results. However it maybe sometimes preferable to use a simpler tool with a better integration with your desktop interfaces. Several solutions exist:
- The Recoll KIO module allows starting queries and viewing results from the Konqueror browser or KDE applications Open dialogs.
- The recollrunner krunner module allows integrating Recoll search results into a krunner query.
- The Ubuntu Unity Recoll Lens lets you access Recoll search from the Unity Dash. More info here.
- The Recoll Web UI lets you query a Recoll index from a web browser
Recoll also has Python and PHP modules which can allow easy integration with web or other applications.
Stemming is a process which transforms inflected words into their most basic form. For example, flooring, floors, floored would probably all be transformed to floor by a stemmer for the English language.
In many search engines, the stemming process occurs during indexing. The index will only contain the stemmed form of words, with exceptions for terms which are detected as being probably proper nouns (ie: capitalized). At query time, the terms entered by the user are stemmed, then matched against the index.
This process results into a smaller index, but it has the grave inconvenient of irrevocably losing information during indexing.
Recoll works in a different way. No stemming is performed at query time, so that all information gets into the index. The resulting index is bigger, but most people probably don't care much about this nowadays, because they have a 100Gb disk 95% full of binary data which does not get indexed.
At the end of an indexing pass, Recoll builds one or several stemming dictionaries, where all word stems are listed in correspondence to the list of their derivatives.
At query time, by default, user-entered terms are stemmed, then matched against the stem database, and the query is expanded to include all derivatives. This will yield search results analogous to those obtained by a classical engine. The benefits of this approach is that stem expansion can be controlled instantly at query time in several ways:
- It can be selectively turned-off for any query term by capitalizing it (Floor).
- The stemming language (ie: english, french...) can be selected (this supposes that several stemming databases have been built, which can be configured as part of the indexing, or done later, in a reasonably fast way).